(Rare special circumstances may alter that requirement: as when seemingly reliable oral history later indicates that the early primary documents were deliberately written with false data.)
Many studies have shown they don't.
For doing this, scientists get Nobels --- high school kids just get expelled
Many scientists liberally sprinkle references to primary reports that they haven't seen or read in their scholarly articles .
The studies dramatically catch them out when a reference with a crucial spelling or numbers error is traced back to a single scholar who did read the primary document.
He or she then miss-spelled or misquoted some of the information.
The reference error was blindly repeated by a long chain of scholars who simply cut and pasted that reference into their hefty tome ---- for added vitamins and mineral count, no doubt.
Ouch ! Gotcha !
But scientists can at least plea that they do other useful things.
Not so historians who can't or won't read primary documents before reporting on them in their scholarly articles and books.
Because first getting dates and chronologies right is what society pays them the big buck for. More about this in a later post.
A universal and eternal now
Science tends to exist in a sort of eternal and universal now.
A 'true' scientist is genuinely puzzled when someone asks them exactly when and where their historic experiment occurred .
For the scientist makes no mention of date or place in the first, announcing, article.
The scientist usually says this process that they discovered has being going on everywhere in the universe since time began because the laws of nature are eternal and universal.
( Spoken like a true devotee of Modernity ! )
If pushed further they say they first published their work in a peer reviewed journal on (say) January 10th 1928, a good week before the other scientists also claiming priority first published their article.
That then - bizarrely enough - becomes to scientists when something is first discovered - when it first happened.
So HGT (the horizontal gene transfer of DNA) , the biggest scientific event of the last quarter of the twentieth century is said to have been first discovered in January 1928 when Frederick Griffith published his accidental discovery of it.
I think that he actually discovered it five years earlier and won't never have published on it ever, if left alone .
But then a visiting German friend half-jokingly threatened that he would go home, repeat the experiment and then publish it.
A threat made mostly to smoke out the reluctant and disbelieving-of-his-own-eyes Griffith.
If the German had, it would have then gotten Griffith into serious trouble with his government employers for withholding an important scientific discovery, made on their dime !
So, unfortunately, we will ever never know when the 'DNA is the physical basis of genetics' saga actually began.
When was penicillin first injected into a patient
Ask the taxpayer/consumer/patient/punter/ paying customer to give a one word description of what makes an antibiotic different from something like an aspirin and they will all say its a "lifesaver".
But to save lives, antibiotics must be injected or swallowed - not dabbed on the outside skin : antiseptics are useful preventives, but they don't save those already dying of infection.
So the historical beginning of the Age of Antibiotics - to the 99.9999999% of us who don't edit scientific journals and assign scientific priorities by whenever they decide to publish articles - began when the first needle of penicillin was injected into a sick patient with the intention of curing them.
I said patient and I meant it - I don't think the date when penicillin was first injected into healthy volunteers is anywhere near as important.
Some scientists and historians, perhaps not thinking very deeply, don't agree.
I can only say in defence of the general consensus on this subject that penicillin is famous for being discovered as a potent and totally safe lifesaver a good 12 years earlier but had then lain fallow without ever been injected into a dying patient.
Crossing that particular medical Rubicon clearly meant something to the world's doctors of that day.
First though - how do we ever know when something as happened ?
We might say that we had been there when it happened - but we might also be biased and re-assembled what actually happened into what we wanted to happen or feared had happened.
More objective ( for no one is ever totally objective) third parties eye witnesses can be rare.
False medical patient charts
In the case of the first penicillin injections, surely then we can trust hospital records.
But we can't.
Howard Florey's team, for example, never even recorded on her medical charts the penicillin injection they gave "volunteer" Mrs Elva Akers in January 1941 !
They similarly didn't record her name on their own team records - just the date and that the penicillin caused a sharp fever spike in the patient - facts they reported in their major penicillin article in Lancet in August 1941.
(Only a much later search through all the ward records for middle aged female patients in mid January 1941 led her doctors to the one lady (Akers) who had a sharp fever spike on her chart --- presumably in response to the remembered fever inducing impurities in the penicillin.)
Medical staff get into serious career ending trouble for changing patient charts - but many things happen during the course of treatment and are not recorded.
Florey et al might get away with such loosey goosey stuff by arguing that this test for human toxicity (which they then believed was the first in history !) wasn't important enough to record on Akers' chart.
But what if this toxicity test had resulted in her death - after all, for 12 long years this wonder drug had lain fallow for that very reason - what then ?
So establishing scientific firsts is very hard - first because scientists often don't record those discoveries very well and because other scientists don't even read (or get right) what little the first group of scientists chose to record and publish ...